{"id":"500bfe0d-95aa-4685-b3b5-e97f9068f109","arxiv_id":"2608.06855","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper constructs and verifies an explicit tangent-category structure on the category of schemes over a base and shows that quasi-separated schemes are classified up to isomorphism by differential-bundle categories.","lead":"This paper spells out, in full detail, how the category of schemes carries a tangent structure whose tangent space at a scheme is the relative tangent scheme built from Kähler differentials and the relative spectrum functor, and it shows that quasi-separated schemes are determined up to isomorphism by their categories of differential bundles.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.2.16 is not proved: axiom 5.1.1(6) is asserted to follow from affine-local verification and gluing, but the equalizer universal property in Sch/S is global and no gluing lemma is supplied.","rationale":"Reading the paper in good faith, the constructions of Section 5.2 are standard and largely carefully set up: the tangent functor via relative spectrum, the adjunction with the infinitesimal thickening W_S, and the Hopf-algebra data are reasonable. The proof of Theorem 5.2.16, however, is not a complete axiom-by-axiom verification. Axioms (3)-(5) involve equalities of morphisms, so an affine-local check is arguably acceptable; axiom (6) is different because it is a universal property. The paper asserts that the equalizer can be glued from affine local checks, but supplies no lemma showing that equalizers of affine X-schemes are compatible with Zariski descent or that arbitrary non-affine test schemes are covered by the verification. The reader's weakest_assumption identified exactly this gluing step for the equalizer condition, so the central concern is already on record. The reconstruction Theorem 6.1.17 is a short consequence of the Cruttwell-Lemay equivalence and Gabriel-Rosenberg; if those external results apply verbatim to the stated class of quasi-separated schemes, that part is sound, though the paper should state the precise hypotheses of [CL23, Theorem 4.27] instead of the varying labels 4.27/4.28. I do not see a demonstrated mathematical counterexample, so the current CONDITIONAL verdict remains appropriate; the missing gluing proof is a reason to require revision rather than to reject the paper outright.","tokens_in":65800,"tokens_out":15507,"duration_ms":174336,"concrete_test":"Prove the missing descent lemma needed for Theorem 5.2.16(6): for every affine open cover {U_i} of an S-scheme X, if the displayed fork in Definition 5.1.1(6) restricts to an equalizer in each Sch/U_i and the restrictions agree on U_i ×_X U_j, then the original fork is an equalizer in Sch/X and is preserved by T^m. The natural proof translates the fork, via Corollary 3.5.11, into an equalizer/coequalizer of quasi-coherent O_X-algebras and then checks the universal property against arbitrary non-affine test schemes mapping into T X. If the lemma is false, an explicit counterexample should be computed, e.g. X = P^1_k with the structures from Definitions 5.2.2-5.2.14, using a non-affine test scheme such as A^2_k \\setminus {0}. This check directly targets the only unproved step in the verification of Theorem 5.2.16.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result, Theorem 5.2.16, is the point at which the paper's argument is least secure. In the axiom-by-axiom proof, items (3)-(6) are dispatched by saying that every map is affine and therefore it suffices to check affine-locally over X and glue. This is not a routine reduction for axiom (6) of Definition 5.1.1: the diagram involving T^2 X, T X, v, 0_X, and (T*p)_X must be an equalizer in the entire category Sch/S, and the universal property must hold for arbitrary test schemes mapping into T X. Universal properties are not local in any evident sense; one needs a descent/gluing lemma for equalizers of affine X-schemes. No such lemma is stated, proved, or even formulated. The issue is not merely cosmetic: the paper's own footnote 27 concedes that Sch/S lacks pushouts and coequalizers in general, so 'glue the equalizer' cannot be treated as a formal consequence of affine-locality. If the local equalizers do not assemble compatibly, the displayed structure fails Definition 5.1.1(6) and Sch/S is not a tangent category. This is a missing argument rather than an observed contradiction, but it is load-bearing because Section 6 and all differential-bundle consequences presuppose the full tangent structure of Theorem 5.2.16.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops, in a largely expository style, the tangent structure on the category Sch/S whose tangent functor is T_{X/S} = Spec_X(Sym_{O_X}(Ω_{X/S})), the relative tangent scheme of Grothendieck. It builds up the needed apparatus in commutative algebra and quasi-coherent sheaves: fibrations of algebras and modules, relative symmetric algebras, the relative spectrum functor, and Kähler differentials. The central result is Theorem 5.2.16, asserting that Sch/S is a tangent category with the stated p, 0, add, ℓ, and c. The paper then constructs a dual tangent structure on the opposite category and, in Section 6, proves that for quasi-separated schemes X and Y there is an equivalence DBun(X) ≃ DBun(Y) if and only if X ≅ Y, using [CL23] and the Gabriel-Rosenberg reconstruction theorem.","tokens_in":66052,"tokens_out":6395,"duration_ms":71897,"significance":"If the central tangent-category theorem is fully established, the paper provides a useful and explicit bridge between tangent category theory and scheme theory, spelling out how the affine model of Example 5.1.13 glues to arbitrary relative schemes. The fibrational and pseudofunctorial perspective on symmetric algebras and the relative spectrum is presented carefully and is likely to be helpful to readers in both communities. The reconstruction theorem, as the paper itself acknowledges, is essentially a folklore consequence of the external results [CL23] and Gabriel-Rosenberg; its value here is organizational rather than as a new geometric invariant. The affine case and the formal dual tangent structure via [CC14, Proposition 5.17] are convincing, and the paper is honest about which ingredients are imported. The main weaknesses are proof gaps in the global descent steps, not questionable mathematical assertions.","major_comments":[{"comment":"The proof of the equalizer condition in Definition 5.1.1(6) is not complete. The text says that because every morphism involved is affine, it suffices to check the equalizer affine-locally over X and then glue. Universal properties in Sch/S are not local in this way: to show that the displayed diagram is an equalizer one must prove a factorization property for arbitrary test schemes, and no descent or gluing lemma for equalizers of affine X-schemes is stated or proved. This is not a cosmetic issue; footnote 27 concedes that Sch/S fails to admit general pushouts and coequalizers, so the gluing step cannot be treated as automatic. The same pattern is used for items (3)–(5), but the equalizer is the sharpest case because it involves a genuine universal property. This missing argument is load-bearing, since Section 6 and the differential-bundle results presuppose the full tangent structure of Theorem 5.2.16.","section":"§5.2, Theorem 5.2.16, item (6)"},{"comment":"The reconstruction theorem depends on two external results whose precise hypotheses are not quoted in the manuscript: [CL23, Theorem 4.27] and the Gabriel-Rosenberg theorem as stated in Theorem 6.1.1. Corollary 6.1.16 claims DBun(X) ≃ QCoh(X)^op for every quasi-separated S-scheme, and Theorem 6.1.17 applies to quasi-separated schemes, but the original sources may require additional finiteness conditions. For example, one should check whether the equivalence of [CL23] holds for a disjoint union of infinitely many copies of Spec k, which is quasi-separated but not quasi-compact. Please state the exact hypotheses of the imported theorems and either prove the needed consequences or restrict the statements to the class for which the cited results are valid. As written, the chain of equivalences in Theorem 6.1.17 is valid only if the cited theorems apply verbatim to the stated class.","section":"§6.1, Theorem 6.1.17 and Corollary 6.1.16"},{"comment":"The proof of the adjunction (−)×_S W_S ⊣ T_{(−)/S} is carried out by choosing affine open covers and then asserting that the local data glue. It does not verify that the local adjunction bijections agree on double overlaps, nor that the resulting map X → T_{Y/S} is independent of the chosen covers. This matters because Corollary 5.2.4, which uses this adjunction to conclude that T is continuous, is used in item (1) of the proof of Theorem 5.2.16, and Proposition 5.3.1 also uses the adjunction. The gap is likely repairable by writing out the compatibility diagrams, but as stated the proof is a sketch rather than a complete verification.","section":"§5.2, Proposition 5.2.3"}],"minor_comments":[{"comment":"There are typos in the abstract: 'bifibration' appears as 'bfibration' and 'quasi-coherent' appears as 'quesicoherent'.","section":"Abstract"},{"comment":"The numbering of the external result from [CL23] is inconsistent: the proof of Theorem 6.1.17 cites [CL23, Theorem 4.27], while the introduction cites [CL23, Theorem 4.28] for the same equivalence; please make the citation consistent.","section":"References and proof of Theorem 6.1.17"},{"comment":"The proof says 'as in Proposition 2.2.4' when referring to the preceding monadicity argument; this should refer to Proposition 2.1.6.","section":"§2.2, proof of Proposition 2.2.4"},{"comment":"The notation T_{X/S} for the tangent scheme and T^n X/S for the nth iterated wide pullback is easy to confuse; please add a sentence that fixes the notation explicitly before Definition 5.2.2.","section":"§5.2, Definition 5.2.2 and Theorem 5.2.16"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the manuscript is a long expository paper with substantial self-citation to unpublished or in-progress works ([CV25], [PV23], [Voo23], [L V25]). I do not see any scientific misconduct in this, but the editor should ask the authors to quote the exact hypotheses of the imported theorems from [CL23] and from the Gabriel-Rosenberg literature. The equalizer gluing gap in Theorem 5.2.16 is the main technical issue; it is fixable in a revision, and I do not see grounds to doubt the underlying statement. The paper's novelty is mostly expository, so the editor should weigh whether the journal's scope is well served by a paper of this length and level of overlap with the cited literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the take. The paper is a serious, mostly expository dictionary between tangent-category formalism and EGA-style scheme theory. Its main value is the patient construction: T_{X/S}=Spec_X Sym Ω, the fibrations, relative spectrum, and Kähler differentials all lined up so a category theorist can see the scheme-theoretic content. The genuinely new items are modest but real: the explicit dual tangent structure on Sch^op/S, and the observation that for quasi-separated schemes, DBun equivalence recovers the scheme, as a direct corollary of Cruttwell–Lemay and Gabriel–Rosenberg. The author is honest that much is folklore and gives detailed references; the self-citations are contextual, not load-bearing.\n\nWhat the paper does well: it actually writes down the structure maps (addition, vertical lift, flip) at the sheaf level, and it gives explicit affine presentations. That is useful reference material for people working at the tangent-category/scheme interface.\n\nThe soft spot is real and load-bearing. Theorem 5.2.16, the claim that Sch/S is a tangent category, is not fully proved. The affine-local reductions in items (3)–(5) are plausible, but item (6), the equalizer axiom for the vertical lift, requires a universal property in the whole category Sch/S. It is not enough to check the equalizer affine-locally and then say \"glue\", especially since the paper itself notes (footnote 27) that Sch/S lacks pushouts and coequalizers in general. What is missing is a descent/gluing lemma for equalizers of affine X-schemes (or a reformulation of the axiom that avoids the global universal property). Without that, Section 6's reconstruction result rests on an unproved central theorem. I don't see a contradiction, and I expect the gap is fillable, but it needs to be filled.\n\nThe reconstruction theorem itself is a two-step consequence of external results; that is fine as a corollary, but the paper should label it as such more clearly and verify that [CL23, Theorem 4.27] applies to all quasi-separated schemes claimed. This is a minor presentation point, not a fatal flaw.\n\nWho this is for: people who want a careful bridge between tangent categories and scheme theory, and who need explicit computations. It deserves a serious referee; the requested revision is clear: supply the missing gluing lemma for the tangent axioms, tighten the framing of the corollaries, and trim the exposition.","headline":"A genuinely useful expository bridge between tangent categories and scheme theory, with two modest new corollaries, but the main tangent-category theorem has a real gluing gap that needs to be fixed before the paper can be fully trusted.","tokens_in":66587,"tokens_out":2353,"would_cite":false,"duration_ms":25784,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14-02","18-02","14A99","14B10","18F40","18F99"],"pacs":[],"model":"deepseek-v4-flash","headline":"The category of schemes over a fixed base is a tangent category—tangent spaces are relative spectra of symmetric algebras of Kähler differentials—and quasi-separated schemes are determined up to isomorphism by their differential-bundle…","keywords":["tangent category","schemes","relative tangent scheme","Kähler differentials","differential bundles","quasi-coherent sheaves","quasi-separated schemes","reconstruction theorem"],"falsifier":"Compute the equalizer in Definition 5.1.1(6) directly on a non-affine glued scheme, for instance $\\mathbb{P}^1$ covered by two affine lines; if the affine-local equalizer diagrams do not assemble into a global equalizer, Theorem 5.2.16 fails. A second test is to search for two non-isomorphic quasi-separated schemes with an equivalence of differential-bundle categories; any such pair would refute Theorem 6.1.17.","tokens_in":65547,"feed_emoji":"📐","tokens_out":16298,"duration_ms":153639,"temperature":0.7,"pith_summary":"This paper establishes that the category of schemes over a base scheme $S$ is a tangent category, a categorical setting in which the tangent bundle and differentiation can be treated abstractly. The tangent scheme of an $S$-scheme $X$ is built as $T_{X/S}=\\operatorname{Spec}_X(\\operatorname{Sym}_{\\mathcal{O}_X}(\\Omega_{X/S}))$: take the sheaf of Kähler differentials, form its symmetric algebra, and apply the relative spectrum over $X$. The paper shows the tangent-category axioms hold by verifying them on affine opens and gluing through Zariski descent. It closes with a reconstruction theorem: for quasi-separated schemes, an equivalence of categories of differential bundles $\\mathbf{DBun}(X)\\simeq\\mathbf{DBun}(Y)$ occurs exactly when $X\\cong Y$, so the bundle category determines the scheme. This matters because it transfers the differential-geometric technology of tangent categories into scheme theory, with quasi-coherent sheaves playing the role that vector bundles play for manifolds.","feed_headline":"Differential bundles classify quasi-separated schemes","feed_subtitle":"Scheme categories carry tangent bundles, and differential bundles recover a scheme up to isomorphism.","key_machinery":"The load-bearing construction is the relative tangent scheme $T_{X/S}=\\operatorname{Spec}_X(\\operatorname{Sym}_{\\mathcal{O}_X}(\\Omega_{X/S}))$: take the sheaf of Kähler differentials of $X$ over $S$, form its symmetric algebra as a quasi-coherent sheaf of $\\mathcal{O}_X$-algebras, and apply the relative spectrum functor, which sends a quasi-coherent sheaf of algebras on $X$ to a scheme affine over $X$. The cocommutative Hopf algebra structure on the symmetric algebra supplies the addition of tangent vectors, while the gluing argument—checking axioms on affine opens and assembling them by Zariski descent—turns a local algebra construction into a global tangent category. For the reconstruction theorem, the key mechanism is the cited equivalence $\\mathbf{DBun}(X)^{\\mathrm{op}}\\simeq\\mathbf{QCoh}(X)$, together with the classical theorem that quasi-coherent sheaves determine a quasi-separated scheme.","core_discovery":"The paper's central assertion is that the category of schemes over a base scheme $S$ is a tangent category, with tangent functor $T_{(-)/S}$ defined by $T_{X/S}=\\operatorname{Spec}_X(\\operatorname{Sym}_{\\mathcal{O}_X}(\\Omega_{X/S}))$. The proof works affine-locally: on affine opens this is the dual-numbers tangent structure on commutative algebras, and the paper shows how the projection, zero section, addition, vertical lift, and canonical flip glue along the relative spectrum and Zariski descent. The second assertion is that for quasi-separated schemes $X$ and $Y$, $X\\cong Y$ if and only if $\\mathbf{DBun}(X)\\simeq\\mathbf{DBun}(Y)$. The forward direction is immediate from functoriality; the reverse direction passes through a cited equivalence between differential bundles and the opposite of quasi-coherent sheaves, together with the classical reconstruction of a quasi-separated scheme from its category of quasi-coherent sheaves.","pith_inferences":["The reconstruction theorem is essentially a translation of two external results; if those results are later extended beyond quasi-separated schemes, the same proof would transfer the differential-bundle invariant to any class of schemes where quasi-coherent sheaves remain a complete invariant.","The author leaves the dual tangent structure on the opposite category as future work; a concrete next test is whether differential bundles there reproduce quasi-coherent sheaves directly, and whether the absence of general pushouts in $\\mathbf{Sch}_{/S}$ obstructs the construction.","Since the equality of differential-bundle categories over quasi-separated schemes is proved from the affineness of bundle projections, one can expect analogous full tangent-subcategory inclusions for other classes closed under affine bundles and tangent powers, such as separated or quasi-compact quasi-separated schemes."],"forward_implications":["For every $S$-scheme $X$, the tangent scheme $T_{X/S}$ is an internal abelian group over $X$, and all the tangent structure maps are affine.","The tangent functor is representable: maps $X\\times_S S[\\varepsilon]\\to Y$ are the same as maps $X\\to T_{Y/S}$, so infinitesimal paths probe tangent vectors.","Quasi-separated schemes are rigid under differential-bundle equivalence: an equivalence $\\mathbf{DBun}(X)\\simeq\\mathbf{DBun}(Y)$ forces an isomorphism $X\\cong Y$.","For quasi-separated $X$, the equality $\\mathbf{DBun}_{\\mathbf{qsSch}}(X)=\\mathbf{DBun}_{\\mathbf{Sch}}(X)$ means the reconstruction invariant is computed inside the full tangent category without extra finiteness hidden in the definition of a bundle.","Slice tangent structures on $\\mathbf{Sch}_{/X}$ assemble pseudofunctorially as the base scheme varies, giving a coherent family of tangent categories parameterized by schemes."],"supporting_citations":[{"why":"Supplies the definition of tangent category and the dual-tangent-structure machinery used to transfer the affine structure to schemes.","marker":"[CC14]"},{"why":"Supplies the definition of differential bundles and their fibrations, which Sections 5 and 6 use throughout.","marker":"[CC18]"},{"why":"Proves the equivalence $\\mathbf{DBun}(X)^{\\mathrm{op}}\\simeq\\mathbf{QCoh}(X)$ for schemes, the bridge used in the reconstruction theorem.","marker":"[CL23]"},{"why":"Provides the reconstruction theorem that equivalence of quasi-coherent sheaf categories forces an isomorphism of quasi-separated schemes.","marker":"[Ros14]"},{"why":"Is the exposition of the reconstruction theorem and Gabriel spectrum that the paper cites as the basis of its Theorem 6.1.1.","marker":"[Bra18]"},{"why":"Provides the relative spectrum functor and the theory of affine morphisms on which the construction of $T_{X/S}$ depends.","marker":"[GD61]"},{"why":"Gives the definition and stability properties of quasi-separated morphisms used to form the strict tangent subcategory $\\mathbf{qsSch}_{/S}$.","marker":"[GD64a]"},{"why":"Supplies the strong tangent morphism result used for the slice-category pseudofunctor over $\\mathbf{Sch}_{/S}$.","marker":"[PV24]"}],"fun_headline_variants":["Differential bundles determine quasi-separated schemes","Scheme recovered from its differential bundles","Differential bundles pin down scheme isomorphism","Differential bundle data recovers the scheme"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the two cited results the paper does not prove—the identification of differential bundles with opposite quasi-coherent sheaves, and the reconstruction of quasi-separated schemes from quasi-coherent sheaves—hold at the stated level of generality; the tangent-category half also assumes the affine-local verification of the axioms glues, especially for the equalizer condition.","fun_headline_variants_meta":{"raw":{"variants":["Differential bundles determine quasi-separated schemes","Scheme recovered from its differential bundles","Differential bundles pin down scheme isomorphism","Differential bundle data recovers the scheme"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000662,"raw_usage":{"total_tokens":3038,"prompt_tokens":971,"completion_tokens":2067,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":587,"completion_tokens_details":{"reasoning_tokens":2014}},"tokens_in":587,"tokens_out":2067,"duration_ms":16943,"temperature":1.0,"reasoning_tokens":2014,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:31:26.763466+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the equalizer in Definition 5.1.1(6) directly on a non-affine glued scheme, for instance $\\mathbb{P}^1$ covered by two affine lines; if the affine-local equalizer diagrams do not assemble into a global equalizer, Theorem 5.2.16 fails. A second test is to search for two non-isomorphic quasi-separated schemes with an equivalence of differential-bundle categories; any such pair would refute Theorem 6.1.17.","supporting_citations":[],"review_version":1}